42 Rules to Lead by, from Jonathan Rosenberg

2026 ChatGPT Summary

Biggest takeaways (high-signal themes)

1) Communication is an execution lever, not a “soft skill”

• Repeat until it lands (“broken record”) because smart teams are still cognitively saturated. (First Round)
• Default to transparency; “power comes from sharing information not hoarding it,” and back claims with data. (First Round)
• Precision matters: leaders’ words get over-interpreted; be crisp, tell stories, listen more than you talk. (First Round)

Mental model: communication reduces “transaction costs” inside the org; the ROI is nonlinear when coordination is the bottleneck.

2) Design culture to resist status and bureaucracy

• De-rank decisions: avoid the “HPPO” (highest-paid person’s opinion); arguments should win on merit and evidence. (First Round)
• Crush bureaucracy: avoid structures that optimize local interests over the company/user. (First Round)
• Small teams outperform in software; org size often adds latency more than output. (First Round)

Mental model: bureaucracies are incentive machines; they create drift away from users unless actively counterweighted.

3) Hiring quality is a flywheel (and bad eggs are negative leverage)

• Hiring is the core system: great people attract great people (compounding). (First Round)
• Committee gates for hires/promotions to reduce single-manager variance. (First Round)
• Seek passion + breadth; avoid narrow specialization (especially in fast-shifting tech). (First Round)
• Purge “bad eggs”—toxic norms spread faster than good ones. (First Round)

Mental model: one bad hire can impose hidden “taxes” across many relationships; the expected cost is multiplicative.

4) Decision-making: set goals that don’t create the wrong fights

• Start with the right goals because goals drive behavior and conflict. (First Round)
• Consensus ≠ unanimity; avoid decision paralysis and the false safety of everyone nodding. (First Round)
• When in doubt, take the customer’s perspective. (First Round)

Mental model: goal-setting is “mechanism design” for your org; you get the behavior you pay for.

5) Innovation: maximize learning rate, not “success rate”

• Creativity can’t be dictated—you can allocate/budget/measure, but not command it. (First Round)
• Prepare to lose to win: leaders shouldn’t prevent failure; they should prevent slow, uninformative failures. (First Round)
• Don’t overuse the kill switch: encourage many ideas; let selection happen. (First Round)
• Postmortems as a norm turn mistakes into shared organizational memory. (First Round)

Mental model: portfolio thinking—innovation is variance; you manage downside via fast feedback loops.

6) Humility is operational, not moral

• Never stop learning; delegate for leverage; mean what you say (hypocrisy destroys trust). (First Round)
• Watch exits (how people leave is diagnostic of character and culture). (First Round)
• Assume you’re not exempt; self-review as a forcing function. (First Round)

“Reviewer / reader” feedback and reactions (what people push back on)

• “Useful, but not novel” framing: getAbstract’s review says the rules “aren’t groundbreaking,” but are strong reminders for leaders. (getAbstract)
• The biggest controversy is the anti-remote stance: a PM reading list calls the article’s advice valuable overall but explicitly says they “absolutely disagree” with the “working from home…” line. (Medium)
• Skeptical practitioner take (LinkedIn comments): one commenter argues Google’s success wasn’t due to mantras/books but “in spite of them,” enabled by massive cashflow, tolerance for mistakes, and weak accountability. (LinkedIn)
• Concrete “this actually happened” anecdote (LinkedIn comments): a former deputy describes an exec process where leaders had to review hiring packets; if they didn’t, the team lost the hire—reinforcing the “hiring is the company” emphasis. (LinkedIn)

Practical synthesis (if you want one “rule of rules”)

Treat the list as a playbook for: (1) increasing truth and information flow, (2) lowering coordination cost, (3) compounding talent quality, and (4) increasing learning speed while controlling downside. (First Round)